Specific Aim 1: Functional importance of gene-environment interactions in the miRNA network in the integrative molecular epidemiology of lung cancer We have examined the 3'UTR of genes related to lung cancer pathways and searched for overlap between miRNA binding sites and SNPs. For those that potentially altered the thermodynamic interaction between mRNAs and miRNAs, we determined whether these SNPs were associated with lung cancer risk. We undertook a genome-wide analysis of SNPs in mature and pre-miRNAs and examined associations between these SNPs and lung cancer. We have integrated demographic and lifestyle exposures and biomarker data available from the case control study. These studies present a vigorous integrated framework within which to consider genetic variation in miRNA networks and its contribution to cancer risk. 1. Hypothesis: 3'UTR SNPs within miRNA binding sites contribute to lung cancer risk Using the NCI-MD case-control study (486 cases/456 population controls), we found three 3'UTR SNPs associated with lung cancer risk in EA. These SNPs were validated in a Mayo Clinic cohort. We focused on CXCR2 and found that rs1126579-T was associated with a reduced risk of lung cancer in EA and a Japanese cohort. We found that the T allele disrupts a binding site for miR516a-3p. We found that both CXCR2 mRNA and protein were increased with the T allele in clinical samples. An eQTL analysis found that MAPK signaling (downstream of CXCR2) was also increased in samples with the T allele. This is consistent with reports that CXCR2 plays a role in p53-dependent senescence. An innovative interaction model showed that the T allele of rs1126579 statistically interacts with circulating levels of IL-8. The results support a model whereby individuals with both the T allele of rs1126579 in CXCR2 and high circulating IL-8 levels have a greater reduced risk of lung cancer than would be expected from the independent effects of either marker alone. We are validating this hypothesis by overexpressing CXCR2 and IL-8 in normal bronchial epithelial cells. These findings suggest that the CXCR2 SNP and circulating levels of serum IL-8 could be used to effectively screen high-risk individuals for screening programs. We received approval to use samples from the NLST cohort to test the hypothesis that circulating levels of IL-8 enhance the sensitivity and specificity of low-dose helical CT for diagnosing lung cancer. We will examine the functional implications between other SNPs identified in this study and lung cancer. Our studies found that the SNP in DRD1 is independent of smoking, therefore we are testing the hypothesis that the association between this SNP and lung cancer risk is modified by exposure to second-hand tobacco smoke. 2. Hypothesis: SNPs within miRNAs contribute to lung cancer risk and survival We conducted a genome-wide search for SNPs in miRNA genes. While such events are rare, one SNP, rs4919510 located within the mature region of mir-608, and was associated with an increased risk of lung cancer in EA, AA, and Japanese populations. We found that the G allele was associated with a reduced risk of mortality. To examine the correlation between this SNP and a lung cancer phenotype, we designed pre-miRNA GFP-tagged constructs representing the C and G allele of rs4919510 and generated cell lines expressing each miRNA. MiR-608 induces apoptosis in both transformed and non-transformed cell lines and it reduces cell proliferation. However, comparisons between the C and G allele of rs4919510 in the induction of apoptosis or proliferation did not reveal differences. Moreover, a cis eQTL analysis of miR-608 expression did not reveal a significant difference in mature miRNA expression, suggesting that the SNP did not affect pre miRNA processing. However, a genome-wide eQTL analysis stratified by rs4919510 genotypes revealed 2 transcripts differentially expressed in samples with the G allele. The transcripts represented MRPL43 variant 4 and MRPL43 variant 1 isoforms. Interestingly, expression of variant 4 was significantly increased in samples with the G allele, while variant 1 was significantly decreased. We validated this eQTL analysis in Japanese, EA, and AA samples, as the risk association was observed in these three cohorts. MRPL43 is a sub-unit of the large mitochondrial ribosome. Pathway analysis of those pathways globally altered in tumor tissues from the CC, CG and GG genotype tissues demonstrated that mitochondrial dysfunction was the main pathway altered in samples with the G allele, thereby suggesting an association between this SNP, alternative splicing of MRPL43 and mitochondrial function in cancer. We are now examining the expression of the MRPL43 splice variants in lung cancer cell lines that represent three rs4919510 genotypes. Using siRNA to knockdown variant specific isoforms of MRPL43 we will investigate how MRPL43 and its isoforms contribute to mitochondrial dysfunction and lung carcinogenesis. To uncover novel miRNA-related SNPs associated with lung cancer, we are performing deep sequencing of 400 lung cancer cases and 300 population controls in 100 lung cancer associated miRNA genes. Sequencing using the Ion Torrent platform is currently ongoing. After validation, candidate SNPs will be functionally investigated using a similar integrative approach to that outlined above. Specific Aim 2: Identify the genomic facet of health disparities in human cancer 1. Hypothesis: GWAS of African American lung cancer will identify potential functional polymorphisms involved in health disparity Lung cancer incidence and mortality rates are higher among AA than EA. As evidence for potential etiologic differences in lung carcinogenesis amongst ethnic groups, some studies indicated that the molecular signature of lung cancer is different across populations. Given the high mortality rates of lung cancer in AA, more research is needed to investigate potential racial differences in lung cancer as these differences may guide potential targets for prevention or therapy. LHC has initiated the first GWAS of lung cancer in AA. AAs are an admixed population, and the proportion of this admixture can vary significantly across individuals. As our GWAS study includes populations from multiple study sites across the U.S., the development of our admixture mapping tool has been a central resource to the analysis of the GWAS data and allows us to control our association analyses for the degree of admixture in each study site. Genotyping and initial data analysis are complete. We have identified several loci associated with risk of lung cancer previously associated with lung cancer risk, including 15q25.1 and 5p15.33. We have identified loci not previously associated with lung cancer in either EA or Asians, such as COL2A1 (12q13.11) and ZFR (5p13.3). These SNPs are undergoing validation in a separate cohort, after which we will conduct an imputation analysis on the top hits using 1,000 Genome data to infer the major risk loci. COL2A1 (collagen, type II, alpha 1) codes for the alpha-1 chain of type II collagen. Interestingly COSMIC database records indicate that approximately 6% of lung tumors carry somatic mutations in this gene; but this gene has not previously been functionally related to lung cancer. We will test whether the SNP affects transcription and translation of COL2A1 using tissue samples from AA lung cancer in the NCI-MD Cohort entered into the GWAS study. We will evaluate the effect of rs17863783 on splicing and expression of COL2A1. Further in vitro and in vivo experiments will then be designed based on these preliminary findings to test the function of COL2A1 in lung carcinogenesis. The study will be used to assess the contribution of loci to specific histologic types of lung cancer.